Data analysis method and system based on three-dimensional map display

Abstract

The invention belongs to the technical field of data processing, and discloses a data analysis method and system based on three-dimensional map display, and the method comprises the steps: collecting real-time data from a plurality of data sources, adding geographic position information to each piece of data after preprocessing, marking the spatial position of the data in a three-dimensional map through geographic coordinates, and obtaining a data analysis result; a standardized three-dimensional map basic model is loaded and constructed, a three-dimensional space environment is constructed through a GIS standard, different marks are visually marked on a map according to attributes and features of data, and a convolutional neural network is used for automatically performing anomaly detection and classification on a time sequence of the data. And generating and displaying a dynamic animation according to the time dimension of the data. According to the invention, the three-dimensional map technology is utilized to enhance the intuition, interactivity and flexibility of data analysis, and meanwhile, the efficiency of data monitoring and exception handling is improved through intelligent data early warning and dispatching command functions.

Description

Technical Field

[0001] The present invention belongs to the field of data processing technology, and specifically relates to a data analysis method and system based on three-dimensional map display. Background Art

[0002] With the rapid development of information technology, enterprises and government agencies have accumulated massive amounts of business data in their daily operations. These data cover a variety of business scenarios, such as enterprise production management, marketing, logistics scheduling, urban planning, etc. Traditional data analysis tools are mainly based on two-dimensional tables and report displays, presenting data through bar charts, pie charts, etc., but this type of display has the problem of insufficient intuitiveness and interactivity, especially in data analysis scenarios that require spatial location association and dynamic monitoring, it is difficult to meet actual needs.

[0003] Traditional data analysis methods mostly rely on two-dimensional reports and graphical displays, such as Excel, PowerBI and other tools, which generate fixed analysis reports by calculating and processing raw data. However, two-dimensional reports are difficult to effectively present the spatial distribution of data, and the spatial dimension is missing. Especially in scenarios involving geographic information, such as logistics route optimization and urban facility management, the limitations of traditional charts are particularly obvious. It is impossible to intuitively present the changing trends and distribution patterns of data in geographic space. Moreover, traditional reports generally display data in a static form. Users lack flexible dynamic interaction methods during the analysis process and cannot adjust the dimensions and levels of data display according to real-time needs, resulting in insufficient dynamic interactivity.

[0004] In order to solve the problem of missing spatial dimensions in data analysis, geographic information systems (GIS) and three-dimensional map technology have gradually been applied to the field of data visualization in recent years. GIS technology can associate data with its geographical location by processing and displaying geographic coordinate data, thereby realizing the visualization of data in space.

[0005] Traditional GIS is mostly used for static map display. With the development of 3D visualization technology, visualization based on 3D maps has gradually become a new trend. 3D maps can not only display elements such as buildings and roads in physical space, but also realize dynamic data mounting and real-time display. However, the existing data display technology based on 3D maps is mainly focused on the static display level, and it is difficult to realize complex dynamic data analysis, especially in scenarios such as enterprise management and urban operations, where users need to analyze data in real time and make decisions based on the analysis results. In addition, most existing 3D map display systems lack intelligent early warning and command and dispatch functions, and are unable to respond quickly and take effective measures when data anomalies or emergencies occur.

[0006] In actual application scenarios such as enterprise and city management, timely data warning and dispatch command are crucial. For example, in a smart factory, the real-time operating status of production line equipment needs to be monitored at all times. When an abnormality occurs in the equipment, the system should be able to issue an early warning immediately and assign relevant personnel to handle the fault through the dispatch system. In urban management, abnormal situations in systems such as transportation and public safety also need to be monitored in a timely manner through the early warning system, and effectively commanded and dispatched through the dispatch center.

[0007] Therefore, how to combine three-dimensional map display technology with data warning and dispatch functions to form a comprehensive system integrating data analysis, visualization, warning, and command and dispatch has become a major challenge in the current field of data analysis and display. Summary of the invention

[0008] In view of the above problems, the present invention provides a data analysis method and system based on three-dimensional map display, which aims to solve the shortcomings of traditional two-dimensional data analysis in spatial display, dynamic interaction, real-time data monitoring and exception handling, and use three-dimensional map technology to enhance the intuitiveness, interactivity and flexibility of data analysis. At the same time, through intelligent data early warning and dispatch command functions, the efficiency of data monitoring and exception handling is improved. The system can be widely used in scenarios such as enterprise management, intelligent manufacturing, and urban operations, providing users with full-process services from data collection and processing to visual display, early warning and dispatch, helping users to achieve accurate and efficient data analysis and decision support.

[0009] To achieve the above object, the present invention provides the following technical solutions.

[0010] The data analysis method based on three-dimensional map display of the present invention comprises the following steps: S10, collect real-time data from multiple data sources, and introduce graph neural network for data fusion and preprocessing; S20, adding geographic location information to each piece of preprocessed data, and marking the spatial location of the data in the three-dimensional map by geographic coordinates; S30, load and build a standardized 3D map basic model, build a 3D space environment through GIS standards, dynamically customize the 3D map model according to the application scenario, and support users to add and adjust different types of data layers on multi-level maps; S40, spatially associate the pre-processed data with the three-dimensional map using geographic coordinates, use visualization to mark different markers on the map according to the attributes and characteristics of the data, support custom markers and layer overlays, and support display of specific data levels; S50, using convolutional neural networks to automatically detect and classify anomalies in the time series of data, and generating and displaying dynamic animations based on the time dimension of the data; S60, monitors data changes in real time through the built-in intelligent data analysis engine, and issues data warnings based on abnormal patterns analyzed from historical data and preset warning rules.

[0011] As a further solution of the present invention, the data source includes Internet of Things devices (IoT sensors, such as temperature sensors and humidity sensors), mobile devices (GPS locations, accelerometers), third-party API interfaces (weather data, traffic data), and databases (historical records, user behaviors).

[0012] As a further solution of the present invention, when real-time data is collected from multiple data sources, a communication interface with the data source is constructed through the MQTT protocol, and high-throughput real-time data stream processing is performed through the Kafka collection framework. The real-time collected data is processed and buffered in real time through the message queue.

[0013] As a further solution of the present invention, in S10, when a graph neural network is introduced for data fusion and preprocessing, the following steps are included: The collected multi-source data are regarded as nodes in the graph neural network, and the edges in the graph neural network are constructed according to the relationship between the data. A weight is assigned to each edge to indicate the strength of the data association. Process graph structure data through graph attention network, and use the adjacency matrix of graph neural network to aggregate and update node features; The graph convolution operation is applied to the data of each node, the data features of the surrounding neighboring nodes are weighted summed, and the node data is fused; Through graph neural network preprocessing, the preprocessed data generates a feature vector containing fused multi-dimensional information for each node.

[0014] As a further solution of the present invention, in S20, marking the spatial position of the data in the three-dimensional map by geographic coordinates includes the following steps: According to the geographical location information added to each piece of data, geocoding is used to convert the address and place name of each piece of data into longitude and latitude coordinates; Convert latitude and longitude coordinates to standardized geographic coordinates and associate them with data entries; The standardized coordinate information of each piece of data is stored as a data structure containing geographic location information; Use GIS to load the 3D map model, convert the geographic coordinates into the spatial position in the 3D map system, and map the 2D longitude and latitude information to the XYZ coordinates in the 3D model; Use the map engine API to create marker points for each geographic coordinate on the 3D map to mark the spatial location; Customize labeling and hierarchical display based on data attributes and features.

[0015] As a further solution of the present invention, S30 is specifically: loading the three-dimensional map basic model, loading the underlying geographic data through a standardized geographic data format, using a coordinate reference system to perform coordinate conversion, loading geographic data, using terrain data to build a three-dimensional terrain model, loading three-dimensional buildings and city models, building a three-dimensional scene, using tile slicing to divide the geographic data into tiles of different levels, and loading corresponding hierarchical data according to the view zoom level when the user interacts.

[0016] As a further solution of the present invention, in S40, different marks are marked on the map by visualization according to the attributes and characteristics of the data, including the following steps: Classify the preprocessed data and determine the data points of different categories; Define the color, shape and size of the marker style according to different attributes of the data; Dynamically set the marker style on the map based on the characteristics of each piece of data; In the 3D GIS engine, you can customize the marking style through the API and apply it to different categories of geographic objects, and dynamically adjust the marking style according to real-time data changes.

[0017] As a further solution of the present invention, in S40, when the custom mark and the layer are superimposed, it includes: Allow users to customize the marker content on the map. The interface allows users to upload custom icons and apply them as markers on the map based on the icon's URL; The map is divided into multiple layers including terrain layers, building layers and marker layers, and users are supported to dynamically turn on or off different layers. Each layer is overlaid and sorted according to user needs. Overlay and display multi-layer data through the layer management function of the GIS engine; Providing a layer management interface allows users to choose to display or hide specific layers through an interactive interface.

[0018] As a further solution of the present invention, in S50, a convolutional neural network is used to automatically detect and classify anomalies in the time series of data, including the following steps: Get time series data and use sliding windows to split the time series data into subsequences of fixed length; Construct and train a convolutional neural network, set a threshold based on the training results, determine whether a data point is abnormal, and classify the data point as normal or abnormal for anomaly detection; Use the Matplotlib library to generate dynamic animations, visualize time series data and detection results, and dynamically display anomalies.

[0019] As a further solution of the present invention, in S60, when performing data warning based on abnormal patterns analyzed from historical data and preset warning rules, it also includes using three times the standard deviation to perform real-time data monitoring and abnormality detection, and triggering warnings in combination with preset business rules.

[0020] As a further solution of the present invention, when three times the standard deviation is used for real-time data monitoring and anomaly detection, the following steps are included: Determine the sliding window size for calculation and obtain the sample window; calculate the mean and standard deviation of historical data; Define a threshold interval. When a new data point falls outside the threshold interval, the data point is considered abnormal. Each newly arrived real-time data point is tested to determine whether it falls within the threshold range. According to the test results, if the data point falls outside the threshold range, it is judged as abnormal, triggering an abnormal alarm, recording the abnormal data point and timestamp, and saving it in the database; When new data points arrive, the sliding window is updated and the mean and standard deviation are recalculated.

[0021] The present invention also provides a data analysis system based on three-dimensional map display, comprising: Data collection module: used to collect real-time data from multiple data sources and introduce graph neural networks for data fusion and preprocessing; Geographic information tagging module: used to add geographic location information to the preprocessed data, associate the data with the physical location through geographic coordinates, and mark the spatial location on the three-dimensional map; 3D map construction module: used to load and build standardized 3D map models, build 3D space environment using GIS (Geographic Information System) standards, dynamically customize map models according to different application scenarios, and support multi-level map views; Layer management module: allows users to add, adjust and manage different data layers on the 3D map; Data visualization module: used to display data on a 3D map using different marking forms based on its attributes and features; Anomaly detection module: used to automatically analyze and detect anomalies in the time series of data based on convolutional neural networks (CNN). The system can classify data according to the law of data changes and identify abnormal trends; Dynamic data display module: used to generate dynamic animation according to the time dimension, and dynamically display the time series data on the three-dimensional map in the form of animation; Early warning module: When data anomalies are triggered, the system automatically sends notifications and alarms according to preset early warning rules.

[0022] Compared with the prior art, the data analysis method and system based on three-dimensional map display proposed by the present invention have the following beneficial effects: 1. Improved accuracy of data fusion and processing. The present invention fuses multi-source data through graph neural networks. The system can process complex heterogeneous data, improve the correlation and consistency between data, and effectively reduce redundant data and noise. The improved accuracy of data fusion makes the system more reliable in data analysis, laying a solid foundation for anomaly detection and decision support; 2. Support accurate geographic spatial association. In the present invention, each piece of data is accompanied by geographic coordinate information, so that the data can be accurately associated with the physical space location on the three-dimensional map. In different application scenarios, the data can be intuitively presented in the corresponding three-dimensional geographical environment, which is suitable for multiple fields such as urban management, logistics monitoring, and environmental monitoring, and improves the operability of data analysis; 3. Dynamic 3D map display and visualization are realized. The present invention adopts GIS standards to construct 3D maps, supports dynamic loading and construction of 3D models in different scenarios, and allows multiple data layers to be superimposed on the map. Users can customize marking styles and layers according to their needs, and intuitively present complex data relationships through 3D visualization technology, facilitating multi-dimensional information interaction; 4. Achieved efficient time series analysis and anomaly detection. The present invention uses convolutional neural networks to analyze time series data, realize automatic classification and anomaly detection, and effectively improve the speed and accuracy of anomaly detection. Through the analysis and pattern recognition of historical data, the system can detect potential risks in advance and respond to abnormal events in real time; 5. Real-time monitoring and intelligent early warning are realized. The system has a built-in intelligent data analysis engine, which can monitor data changes in real time and issue abnormal early warnings based on historical data patterns and preset rules. The system supports multiple early warning mechanisms, such as SMS, email and system notifications, to ensure that users receive risk warnings in a timely manner and improve response efficiency; 6. Dynamic data display and animation enhance user experience. The system supports displaying the time series changes of data in the form of animation on a three-dimensional map, enhancing the user's intuitive perception of data changes. The present invention improves the user's experience in the data analysis and decision-making process through dynamic display, and is suitable for scenarios such as traffic flow monitoring, meteorological data display, and logistics visualization; 7. Support flexible multi-level view management. The system of the present invention provides a multi-level map management module, which supports users to view and analyze data at different levels, and realizes multi-dimensional analysis from macro to micro. Flexible view management enables the system to meet the diverse needs of users in different fields, such as emergency response and commercial data analysis.

[0023] In summary, the present invention combines map neural network, convolutional neural network, GIS three-dimensional map and dynamic visualization technology to achieve accurate data collection, fusion, analysis and display. The system not only improves the efficiency and accuracy of data processing, but also supports real-time monitoring and early warning, greatly enhancing user experience and data insight. The system design with strong adaptability and good scalability lays the foundation for its application in smart cities, logistics management, environmental monitoring and other fields. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 A flowchart of a data analysis method based on three-dimensional map display according to an embodiment of the present invention; Figure 2 This is a flow chart of introducing a graph neural network for data fusion and preprocessing in a data analysis method based on three-dimensional map display in an embodiment of the present invention; Figure 3 A flowchart of marking the spatial position of data in a three-dimensional map in a data analysis method based on three-dimensional map display in an embodiment of the present invention; Figure 4 A flowchart marked on a map in a data analysis method based on three-dimensional map display in an embodiment of the present invention; Figure 5 A flowchart of anomaly detection and classification in a data analysis method based on three-dimensional map display in an embodiment of the present invention; Figure 6 The present invention is a flowchart of a data analysis method based on three-dimensional map display in an embodiment of the present invention, which uses three times the standard deviation to perform real-time data monitoring and anomaly detection. DETAILED DESCRIPTION

[0025] The present invention is further described below in conjunction with the accompanying drawings and specific implementation methods. It should be noted that, under the premise of no conflict, the various embodiments or technical features described below can be arbitrarily combined to form new embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0026] It should be noted that all expressions using "first" and "second" in the embodiments of the present invention are intended to distinguish two non-identical entities or non-identical parameters with the same name. It can be seen that "first" and "second" are only for the convenience of expression and should not be understood as limitations on the embodiments of the present invention. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, other steps or units inherent to a process, method, system, product or device that includes a series of steps or units.

[0027] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0028] The flowcharts shown in the accompanying drawings are only examples and do not necessarily include all the contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps may also be decomposed, combined or partially merged, so the actual execution order may change according to actual conditions.

[0029] Some embodiments of the present invention are described in detail below in conjunction with the accompanying drawings. In the absence of conflict, the following embodiments and features in the embodiments can be combined with each other.

[0030] Aiming at the problems of traditional two-dimensional data analysis in spatial display, dynamic interaction, real-time data monitoring and exception handling. The present invention proposes a data analysis method and system based on three-dimensional map display, which uses three-dimensional map technology to enhance the intuitiveness, interactivity and flexibility of data analysis, and improves the efficiency of data monitoring and exception handling through intelligent data early warning and dispatch command functions. The system can be widely used in scenarios such as enterprise management, intelligent manufacturing, and urban operations, providing users with full-process services from data collection and processing to visual display, early warning and dispatch, helping users to achieve accurate and efficient data analysis and decision support.

[0031] See also Figure 1 As shown, an embodiment of the present invention provides a data analysis method based on three-dimensional map display, the method comprising the following steps:

[0032] S10. Collect real-time data from multiple data sources and introduce graph neural networks for data fusion and preprocessing.

[0033] In this step, the data sources include IoT devices, mobile devices, third-party API interfaces, and databases.

[0034] When collecting real-time data from multiple data sources, a communication interface with the data source is built through the MQTT protocol, and high-throughput real-time data stream processing is performed through the Kafka collection framework. The real-time collected data is processed and buffered in real time through the message queue.

[0035] In this embodiment, see Figure 2 As shown in the figure, when introducing graph neural network for data fusion and preprocessing, the following steps are included: S101, treating the collected multi-source data as nodes in a graph neural network, constructing edges in the graph neural network based on the relationship between the data, and assigning a weight to each edge to indicate the strength of the data association; S102, processing graph structure data through the graph attention network, and aggregating and updating node features using the adjacency matrix of the graph neural network; S103, applying graph convolution operation to the data of each node, performing weighted summation on the data features of surrounding neighboring nodes, and fusing the node data; S104, through graph neural network preprocessing, the preprocessed data generates a feature vector containing fused multi-dimensional information for each node.

[0036] S20, adding geographic location information to each piece of preprocessed data, and marking the spatial location of the data in the three-dimensional map through geographic coordinates.

[0037] In this step, see Figure 3 As shown, marking the spatial position of data in a three-dimensional map by geographic coordinates includes the following steps: S201, according to the geographical location information added to each piece of data, using geocoding to convert the address and place name of each piece of data into longitude and latitude coordinates; S202, converting the longitude and latitude coordinates into standardized geographic coordinates and associating them with the data entry; S203, storing the standardized coordinate information of each piece of data as a data structure containing geographic location information; S204, using GIS to load the three-dimensional map model, converting the geographic coordinates into the spatial position in the three-dimensional map system, and mapping the two-dimensional longitude and latitude information to the XYZ coordinates in the three-dimensional model; S205, using the API of the map engine to create a marking point for each geographic coordinate on the three-dimensional map to mark the spatial position; S206. Customize labeling and hierarchical display based on data attributes and features.

[0038] Taking the environmental monitoring system as an example, the data analysis method in this embodiment is applied to the monitoring and visualization of environmental parameters. Among them, the collected multi-source data includes sensor data, weather station data and environmental monitoring station API, which are processed and preprocessed in real time through the MQTT protocol and Kafka framework. Sensor data (PM2.5, temperature and humidity) are fused and preprocessed through graph neural networks, nodes represent sensors at different locations, and edges represent the spatial correlation between sensors. When geographic location marking and three-dimensional display, longitude and latitude information is added to the monitoring data through geocoding, and it is converted into a marker point in three-dimensional space. The three-dimensional map model is loaded using GIS, and sensor data of different pollution levels are displayed on the map in the form of custom markers. Different pollution levels are distinguished by markers of different colors. The markers of highly polluted areas are automatically enlarged, and a heat map effect is formed on the map. Users can choose to view pollution changes in different time periods and dynamically display historical time series data.

[0039] Through the collection of multi-source data and the integration of graph neural networks, the system can efficiently process complex heterogeneous data and generate multi-dimensional information. At the same time, GIS technology is used to visualize the data, which enhances the user's intuitive perception of spatial data. Through multi-level visualization and dynamic time series data analysis, the accuracy and timeliness of data analysis are improved. It is suitable for many fields such as environmental monitoring, urban management, and logistics tracking.

[0040] S30. Load and build a standardized 3D map basic model, construct a 3D space environment through GIS standards, dynamically customize the 3D map model according to the application scenario, and support users to add and adjust different types of data layers on multi-level maps.

[0041] In this step, when loading and building a standardized three-dimensional map base model, the three-dimensional map base model is loaded, the underlying geographic data is loaded through a standardized geographic data format, coordinate conversion is performed using a coordinate reference system, geographic data is loaded, a three-dimensional terrain model is built using terrain data, three-dimensional buildings and city models are loaded, a three-dimensional scene is built, and tile slicing is used to divide the geographic data into tiles of different levels. When the user interacts, the corresponding level data is loaded according to the view zoom level.

[0042] In this embodiment, when loading the 3D map basic model, the GIS standard is used to construct the 3D space environment, and when loading the 3D map basic model, standardized geographic data formats such as Shapefile, GeoJSON, WMS, etc. are used. Coordinate conversion is performed through the coordinate reference system (CRS) to ensure accurate mapping between different coordinate systems, and the 3D terrain model is constructed using terrain data (such as DEM data), and the underlying geographic data, including terrain, rivers, roads and other infrastructure, is loaded.

[0043] When constructing three-dimensional building and city models, use city building models, such as three-dimensional building models in CityGML format, map two-dimensional geographic data into three-dimensional buildings and urban landscapes, slice the geographic data into tiles, and divide it into tile data of different levels. When the user interacts, the data of the corresponding level is dynamically loaded according to the view zoom level, thereby improving loading efficiency and performance.

[0044] When building a three-dimensional scene, the three-dimensional map model is dynamically customized according to the application scenario requirements to achieve map display in a specific scenario. On this basis, users can add and adjust different types of data layers. Multi-level map display is supported. Users can switch three-dimensional views according to different perspectives, providing a full range of spatial display from terrain, buildings to environmental data.

[0045] S40. Use geographic coordinates to spatially associate the preprocessed data with the three-dimensional map, use visualization to mark different markers on the map according to the attributes and characteristics of the data, support custom markers and layer overlays, and support the display of specific data levels.

[0046] In this step, the data preprocessed by the graph neural network is associated with the three-dimensional map using geographic coordinates, and each data point is accurately located at the corresponding position in the three-dimensional space. Different marking styles are created on the three-dimensional map through data features and attributes, such as color, shape, and size to distinguish different categories of data.

[0047] See also Figure 4 As shown, different markers are marked on the map using visualization according to the attributes and characteristics of the data, including the following steps: S401, classifying the preprocessed data to determine data points of different categories; S403, defining the color, shape and size of the marking style according to different attributes of the data; S403, dynamically setting the marking style on the map according to the characteristics of each piece of data; S404. In the 3D GIS engine, a marking style is customized through an API and applied to different categories of geographic objects, and the marking style is dynamically adjusted according to real-time data changes.

[0048] Among them, in the data classification and labeling definition, the pre-processed data is classified, and the attributes of each data point (such as category, time, source, etc.) are determined to prepare for subsequent visualization. According to the different attributes of the data, the color, shape, and size of the labeling style are defined. For example, the pollution source in the environmental monitoring data can be marked with a red circle, while the meteorological data is marked with a blue square. According to the characteristics of each piece of data, the style of the label on the map is dynamically adjusted to ensure that the data is displayed intuitively and clearly on the map. On the three-dimensional map, the system can dynamically update the labeling style according to the changes in real-time data. For example, when the weather changes, the labeling color of the meteorological data will change with the changes in meteorological conditions. The API interface supports users to upload custom icons. Users can use the uploaded icon URL as a label to apply it to different geographical objects on the map (such as buildings, transportation stations, sensor locations, etc.). The layer management function supports the superposition display of multi-level data, dividing the map into multiple layers (such as terrain layer, building layer, label layer, etc.). Users can dynamically open or close layers as needed to display and manage data in layers. The system provides a layer management interface, allowing users to choose to display or hide specific layers through the interactive interface. For example, users can choose to display only the marker layer or the building layer, thereby customizing the display according to their needs.

[0049] In this embodiment, when the custom mark and the layer are superimposed, it includes: Allow users to customize the marker content on the map. The interface allows users to upload custom icons and apply them as markers on the map based on the icon's URL; The map is divided into multiple layers including terrain layers, building layers and marker layers, and users are supported to dynamically turn on or off different layers. Each layer is overlaid and sorted according to user needs. Overlay and display multi-layer data through the layer management function of the GIS engine; Providing a layer management interface allows users to choose to display or hide specific layers through an interactive interface.

[0050] For example, in a scenario of a city traffic monitoring system, the application of this embodiment is as follows: (1) 3D map basic model: Load the 3D map model of the city through GIS technology to display the city’s roads, buildings, bridges and other important facilities, and construct the terrain through DEM data; (2) Data classification and labeling: Real-time data is collected from multiple sources such as traffic monitoring cameras, GPS devices, and traffic flow sensors. The system classifies different traffic data, including: Traffic accident points: marked in red; Road congestion areas: marked in orange; Vehicle real-time location information: marked in blue; (3) Spatial association and dynamic display: Based on the geographic location information in the data, the traffic accident points, congestion areas, etc. are accurately located in the three-dimensional map, and the marking style is automatically updated according to the real-time data changes. For example, when a new traffic accident occurs, the corresponding location on the map will be marked with a red mark, and the size and color of the mark will be automatically adjusted according to the severity of the accident; (4) Customized icons and layer overlays: Transportation departments can upload customized icons to mark specific traffic lights, parking lots, or road closures. Users can use the layer management function to choose to view different data levels, such as choosing to view only the "traffic accident point" or "traffic flow" layer.

[0051] The present invention ensures that geographic data from different data sources can be seamlessly integrated by using GIS standards and coordinate reference systems, optimizes loading performance through tile slicing, supports user-defined marking styles, and realizes dynamic updating and visualization of data on the map in combination with real-time changes in data. It is suitable for multiple fields such as urban management, environmental monitoring, and traffic control.

[0052] S50. Use convolutional neural networks to automatically detect and classify anomalies in the time series of data, and generate and display dynamic animations based on the time dimension of the data.

[0053] In this step, see Figure 5 As shown in the figure, the convolutional neural network is used to automatically detect and classify anomalies in the time series of data, including the following steps: S501, obtaining time series data, and using a sliding window to divide the time series data into subsequences of fixed length; S502, constructing a convolutional neural network and training it, setting a threshold value according to the training result, determining whether a data point is abnormal, and classifying the data point as normal or abnormal for abnormality detection; S503. Use the Matplotlib library to generate dynamic animations, visualize time series data and detection results, and dynamically display abnormal points.

[0054] Among them, when obtaining time series data, the original time series data is obtained from data sources such as sensors, databases, and logs, and the sliding window method is used to divide the time series data into subsequences of fixed length, such as each subsequence length is 20 data points and the sliding step is 5 data points. When constructing a convolutional neural network (CNN) model, it includes an input layer, a convolution layer, a pooling layer, a fully connected layer, and an output layer. The network can use 1D convolution operations to extract local features of time series data, use normal data and abnormal data for supervised training, use the cross entropy loss function for model optimization, set the classification threshold according to the training results, and judge whether the input data is abnormal. When the predicted abnormal probability is higher than the set threshold, the data point is judged to be abnormal.

[0055] When generating and displaying dynamic animations, the Matplotlib library is used to generate dynamic animations of time series. The animation displays the time series curve of the data, and the location information of the abnormal points is marked. The data curve is updated in real time in the animation, the detected abnormal points are marked, and the classification results of the abnormal points are dynamically displayed (such as red for abnormality and green for normality).

[0056] S60 monitors data changes in real time through the built-in intelligent data analysis engine, and issues data warnings based on abnormal patterns analyzed from historical data and preset warning rules.

[0057] In this step, when issuing data warnings based on abnormal patterns in historical data analysis and preset warning rules, it also includes using three times the standard deviation for real-time data monitoring and anomaly detection, and triggering warnings in combination with preset business rules.

[0058] Among them, see Figure 5 As shown in the figure, when using triple standard deviation for real-time data monitoring and anomaly detection, the following steps are included: S601, determine the sliding window size for calculation, and obtain a sample window; calculate the mean and standard deviation of historical data; S602, defining a threshold interval, and when a new data point falls outside the threshold interval, the data point is considered abnormal; S603, detect each newly arrived real-time data point to determine whether it falls within the threshold range. According to the detection result, if the data point falls outside the threshold range, it is determined to be abnormal, triggering an abnormal alarm, recording the abnormal data point and timestamp, and saving it in the database; S604: When a new data point arrives, update the sliding window and recalculate the mean and standard deviation.

[0059] When analyzing historical data, we can extract typical abnormal patterns (such as periodic fluctuations, sudden abnormalities, etc.) by statistically analyzing historical time series data, establish an abnormal pattern library, and set specific warning rules based on business needs, such as triggering an alarm when the abnormal frequency exceeds a certain number of times. When using the triple standard deviation method for real-time data monitoring, determine the size of the sliding window based on the specific application scenario and data characteristics. For example, use the past 100 data points as the length of the sliding window, calculate the mean and standard deviation of the data points in the sliding window, and obtain the basic statistical parameters for real-time monitoring.

[0060] When setting the threshold interval, it is defined as the range of [mean - 3 times standard deviation, mean + 3 times standard deviation]. When a new data point falls outside this interval, it is considered an anomaly. During real-time detection and abnormal alarm, for each newly arrived real-time data point, determine whether it falls within the threshold interval. If the data point falls outside the interval, it is determined to be abnormal, and the abnormal data point and its timestamp are recorded, the alarm information is sent to the system administrator, and the abnormal point information is saved in the database. When updating the sliding window and recalculating the parameters, whenever a new data point arrives, update the sliding window, remove the oldest data point, add the latest data point, and recalculate the data mean and standard deviation in the window.

[0061] Exemplarily, in a monitoring system for industrial manufacturing equipment, the application of this embodiment is as follows: (1) Time series data acquisition: The time series data of vibration amplitude is collected from the vibration sensor of the production equipment. When the equipment is operating normally, the data has a limited range of variation; when the equipment is abnormal, the vibration data will fluctuate abnormally; (2) Convolutional neural network detection: A convolutional neural network model is used to analyze the vibration data of the equipment in real time. Normal vibration data and vibration data of some known faults are used as samples for model training. When abnormal vibration is detected, the system automatically marks the data point as abnormal and displays an alarm message on the monitoring interface; (3) Triple standard deviation monitoring: Use historical vibration data to calculate the mean and standard deviation, and set [mean - 3 times standard deviation, mean + 3 times standard deviation] as the normal vibration range. When the real-time vibration data exceeds this range, an alarm is triggered immediately and the abnormal information is recorded in the monitoring system; (4) Dynamic display and early warning: The time series curve of vibration data is displayed through Matplotlib dynamic animation, and the mark of data points (normal / abnormal) is updated in real time. The system automatically identifies and warns of potential equipment failures based on the pattern of historical fault data and the preset alarm rules.

[0062] The present invention adopts convolutional neural network to automatically classify time series data, which can quickly identify abnormal data points. It is suitable for fields such as industrial equipment monitoring and financial time series analysis. The method of triple standard deviation is combined with the early warning rules preset by the business to realize real-time monitoring and abnormal alarm of data. The time series data is displayed in the form of dynamic animation, which intuitively reflects data changes and abnormal points, and enhances the visualization effect and user experience of the monitoring system.

[0063] The data analysis method based on three-dimensional map display of the present invention fuses multi-source data through graph neural network. The system can process complex heterogeneous data, improve the correlation and consistency between data, and effectively reduce redundant data and noise. The improved accuracy of data fusion enables the system to have higher reliability in data analysis, laying a solid foundation for anomaly detection and decision support. In the present invention, each piece of data is accompanied by geographic coordinate information, so that the data can be accurately associated with the physical space location on the three-dimensional map. In different application scenarios, the data can be intuitively presented in the corresponding three-dimensional geographical environment, which is suitable for multiple fields such as urban management, logistics monitoring, and environmental monitoring, and improves the operability of data analysis.

[0064] The present invention adopts GIS standards to construct three-dimensional maps, supports dynamic loading and construction of three-dimensional models in different scenarios, and allows multiple data layers to be superimposed on the map. Users can customize marking styles and layers according to needs, and intuitively present complex data relationships through three-dimensional visualization technology, which is convenient for multi-dimensional information interaction. The present invention uses convolutional neural networks to analyze time series data, realize automatic classification and anomaly detection, and effectively improve the speed and accuracy of anomaly detection. Through the analysis and pattern recognition of historical data, the system can detect potential risks in advance and respond to abnormal events in real time. Through the system's built-in intelligent data analysis engine, data changes can be monitored in real time, and abnormal warnings can be issued according to historical data patterns and preset rules. The system supports multiple warning mechanisms, such as text messages, emails, and system notifications, to ensure that users receive risk warnings in a timely manner and improve response efficiency.

[0065] In a second aspect of the embodiments of the present invention, the present invention further provides a data analysis system based on three-dimensional map display, comprising: Data collection module: used to collect real-time data from multiple data sources and introduce graph neural networks for data fusion and preprocessing; Geographic information tagging module: used to add geographic location information to the preprocessed data, associate the data with the physical location through geographic coordinates, and mark the spatial location on the three-dimensional map; 3D map construction module: used to load and build standardized 3D map models, build 3D space environment using GIS (Geographic Information System) standards, dynamically customize map models according to different application scenarios, and support multi-level map views; Layer management module: allows users to add, adjust and manage different data layers on the 3D map; Data visualization module: used to display data on a 3D map using different marking forms based on its attributes and features; Anomaly detection module: used to automatically analyze and detect anomalies in the time series of data based on convolutional neural networks (CNN). The system can classify data according to the law of data changes and identify abnormal trends; Dynamic data display module: used to generate dynamic animation according to the time dimension, and dynamically display the time series data on the three-dimensional map in the form of animation; Early warning module: When data anomalies are triggered, the system automatically sends notifications and alarms according to preset early warning rules.

[0066] Through the above detailed steps and examples, the data analysis system based on three-dimensional map display of the present invention is used to execute the steps of the data analysis method based on three-dimensional map display in the above embodiment, which will not be repeated here. The data analysis system based on three-dimensional map display of the present invention supports displaying the time series changes of data on a three-dimensional map in an animated form, enhancing the user's intuitive perception of data changes. The present invention enhances the user's experience in the data analysis and decision-making process through dynamic display, and is suitable for scenarios such as traffic flow monitoring, meteorological data display, and logistics visualization. The system of the present invention provides a multi-level map management module to support users to view and analyze data at different levels, and realize multi-dimensional analysis from macro to micro. Flexible view management enables the system to meet the diverse needs of users in different fields, such as emergency response and commercial data analysis.

[0067] In summary, the present invention combines map neural network, convolutional neural network, GIS three-dimensional map and dynamic visualization technology to achieve accurate data collection, fusion, analysis and display. The system not only improves the efficiency and accuracy of data processing, but also supports real-time monitoring and early warning, greatly enhancing user experience and data insight. The system design with strong adaptability and good scalability lays the foundation for its application in smart cities, logistics management, environmental monitoring and other fields.

[0068] The above description is only a preferred embodiment of the present invention and is not intended to be a further limitation of the present invention. All equivalent changes made using the contents of the present specification and drawings are within the protection scope of the present invention.

Claims

1. A data analysis method based on three-dimensional map display, characterized in that: The method comprises the following steps: S10, collect real-time data from multiple data sources, and introduce graph neural network for data fusion and preprocessing; S20, adding geographic location information to each piece of preprocessed data, and marking the spatial location of the data in the three-dimensional map by geographic coordinates; S30, load and build a standardized 3D map basic model, build a 3D space environment through GIS standards, dynamically customize the 3D map model according to the application scenario, and support users to add and adjust different types of data layers on multi-level maps; S40, spatially associate the pre-processed data with the three-dimensional map using geographic coordinates, use visualization to mark different markers on the map according to the attributes and characteristics of the data, support custom markers and layer overlays, and support display of specific data levels; S50, using convolutional neural networks to automatically detect and classify anomalies in the time series of data, and generating and displaying dynamic animations based on the time dimension of the data; S60, monitors data changes in real time through the built-in intelligent data analysis engine, and issues data warnings based on abnormal patterns analyzed from historical data and preset warning rules.

2. The data analysis method based on three-dimensional map display according to claim 1, characterized in that: In S10, when real-time data is collected from multiple data sources, a communication interface with the data source is built through the MQTT protocol, and high-throughput real-time data stream processing is performed through the Kafka collection framework. The real-time collected data is processed and buffered in real time through the message queue.

3. The data analysis method based on three-dimensional map display according to claim 2 is characterized in that: In S10, when the graph neural network is introduced for data fusion and preprocessing, the following steps are included: The collected multi-source data are regarded as nodes in the graph neural network, and the edges in the graph neural network are constructed according to the relationship between the data. A weight is assigned to each edge to indicate the strength of the data association. Process graph structure data through graph attention network, and use the adjacency matrix of graph neural network to aggregate and update node features; The graph convolution operation is applied to the data of each node, the data features of the surrounding neighboring nodes are weighted summed, and the node data is fused; Through graph neural network preprocessing, the preprocessed data generates a feature vector containing fused multi-dimensional information for each node.

4. The data analysis method based on three-dimensional map display according to claim 3 is characterized in that: In S20, the spatial position of the data in the three-dimensional map is marked by geographic coordinates, including the following steps: According to the geographical location information added to each piece of data, geocoding is used to convert the address and place name of each piece of data into longitude and latitude coordinates; Convert latitude and longitude coordinates to standardized geographic coordinates and associate them with data entries; The standardized coordinate information of each piece of data is stored as a data structure containing geographic location information; Use GIS to load the 3D map model, convert the geographic coordinates into the spatial position in the 3D map system, and map the 2D longitude and latitude information to the XYZ coordinates in the 3D model; Use the map engine API to create marker points for each geographic coordinate on the 3D map to mark the spatial location; Customize labeling and hierarchical display based on data attributes and features.

5. The data analysis method based on three-dimensional map display according to claim 4 is characterized in that: S30 is specifically as follows: loading the three-dimensional map basic model, loading the underlying geographic data through a standardized geographic data format, using a coordinate reference system for coordinate conversion, loading geographic data, using terrain data to build a three-dimensional terrain model, loading three-dimensional buildings and city models, building a three-dimensional scene, using tile slicing to divide geographic data into tiles of different levels, and loading corresponding hierarchical data according to the view zoom level when the user interacts.

6. The data analysis method based on three-dimensional map display according to claim 5 is characterized in that: In S40, different marks are marked on the map using visualization according to the attributes and characteristics of the data, including the following steps: Classify the preprocessed data and determine the data points of different categories; Define the color, shape and size of the marker style according to different attributes of the data; Dynamically set the marker style on the map based on the characteristics of each piece of data; In the 3D GIS engine, you can customize the marking style through the API and apply it to different categories of geographic objects, and dynamically adjust the marking style based on real-time data changes.

7. The data analysis method based on three-dimensional map display according to claim 6 is characterized in that: In S40, when custom markers and layers are superimposed, the details are as follows: Allow users to customize the marker content on the map. The interface allows users to upload custom icons and apply them as markers on the map based on the icon's URL; The map is divided into multiple layers including terrain layers, building layers and marker layers, and users are supported to dynamically turn on or off different layers. Each layer is overlaid and sorted according to user needs. Use the layer management function of the GIS engine to overlay and display multiple layers of data; Providing a layer management interface allows users to choose to display or hide specific layers through an interactive interface.

8. The data analysis method based on three-dimensional map display according to claim 1, characterized in that: In S50, a convolutional neural network is used to automatically detect and classify anomalies in the time series of data, including the following steps: Get time series data and use sliding windows to split the time series data into subsequences of fixed length; Construct and train a convolutional neural network, set a threshold based on the training results, determine whether a data point is abnormal, and classify the data point as normal or abnormal for anomaly detection; Use the Matplotlib library to generate dynamic animations, visualize time series data and detection results, and dynamically display anomalies.

9. The data analysis method based on three-dimensional map display according to claim 8, characterized in that: In S60, three times the standard deviation is used for real-time data monitoring and anomaly detection, including the following steps: Determine the sliding window size for calculation and obtain the sample window; calculate the mean and standard deviation of historical data; Define a threshold interval. When a new data point falls outside the threshold interval, the data point is considered abnormal. Each newly arrived real-time data point is tested to determine whether it falls within the threshold range. According to the test results, if the data point falls outside the threshold range, it is judged as abnormal, triggering an abnormal alarm, recording the abnormal data point and timestamp, and saving it in the database; When new data points arrive, the sliding window is updated and the mean and standard deviation are recalculated.

10. A data analysis system based on three-dimensional map display, characterized in that: The system is used to execute the data analysis method based on three-dimensional map display according to any one of claims 1 to 9, comprising: Data collection module: used to collect real-time data from multiple data sources and introduce graph neural networks for data fusion and preprocessing; Geographic information tagging module: used to add geographic location information to the preprocessed data, associate the data with the physical location through geographic coordinates, and mark the spatial location on the three-dimensional map; 3D map construction module: used to load and build standardized 3D map models, build 3D space environment using GIS standards, dynamically customize map models according to different application scenarios, and support multi-level map views; Layer management module: allows users to add, adjust and manage different data layers on the 3D map; Data visualization module: used to display data on a 3D map using different marking forms based on its attributes and features; Anomaly detection module: used to automatically analyze and detect anomalies in the time series of data based on convolutional neural networks, classify data according to the law of data changes, and identify abnormal trends; Dynamic data display module: used to generate dynamic animation according to the time dimension, and dynamically display the time series data on the three-dimensional map in the form of animation; Early warning module: When data anomalies are triggered, the system automatically sends notifications and alarms according to preset early warning rules.

Images